Journal of Aging and Physical Activity, 2014, 22, 557-563 http://dx.doi.org/10.1123/JAPA.2012-0241 © 2014 Human Kinetics, Inc.
Official Journal of ICAPA www.JAPA-Journal.com ORIGINAL RESEARCH
Dual-task Performance in Young and Older Adults: Speed-Accuracy Tradeoffs in Choice Responding While Treadmill Walking Phillip D. Tomporowski and Michel Audiffren Thirty-one young (mean age = 20.8 years) and 30 older (mean age = 71.5 years) men and women categorized as physically active (n = 30) or inactive (n = 31) performed an executive processing task while standing, treadmill walking at a preferred pace, and treadmill walking at a faster pace. Dual-task interference was predicted to negatively impact older adults’ cognitive flexibility as measured by an auditory switch task more than younger adults; further, participants’ level of physical activity was predicted to mitigate the relation. For older adults, treadmill walking was accompanied by significantly more rapid response times and reductions in local- and mixed-switch costs. A speed-accuracy tradeoff was observed in which response errors increased linearly as walking speed increased, suggesting that locomotion under dual-task conditions degrades the quality of older adults’ cognitive flexibility. Participants’ level of physical activity did not influence cognitive test performance. Keywords: switch task, cognition, dual-task interference, physical activity, information processing Daily activities require the ability to multitask. Researchers in many fields have long been interested in understanding how people perform multiple tasks concurrently. Numerous studies have shown that dual-task demands lead to degraded performance on one or both tasks (see review by Wickens, 2008). These changes in performance are typically explained in terms of competition for limited resources (Kahneman, 1973). The general notion is that the resources required for processing at any point in time are available only in fixed quantities and that performance deteriorates when resources are exceeded by task demands (Navon & Gopher, 1980). Theorists debate whether resources are best characterized as a single, general purpose entity or as multiple pools (Wickens, 2008). Researchers have investigated dual-task conditions that lead to interference between control of balance or locomotion and cognitive processing (Bock, 2008). The dual-task demands of motoric and cognitive processing have been of particular interest to gerontologists, who posit that age-related declines in sensorimotor processes increase the magnitude of task interference (Li & Lindenberger, 2002; Li, Lindenberger, Freund, & Baltes, 2001; Woollacott & Shumway-Cook, 2002). Older adults consistently have been found to prioritize balance (Shumway-Cook, Baldwin, Polissar, & Gruber, 1997) and locomotor control (Lindenberger, Marsiske, & Baltes, 2000; Rapp, Krampe, & Baltes, 2006) over cognitive performance. Data obtained from these studies, and others, have led some to question the traditional view that balance and locomotion control is primarily reflexive and highly automatized and to suggest that higher-level executive functions play a role in these motoric behaviors (Yogev-Seligmann, Hausdorff, & Giladi, 2008). Regardless of the theoretical view, the allocation of resources to control locomotor behaviors is predicted to reduce available resources for processing competing tasks. Tomporowski is in the Department of Kinesiology at the University of Georgia, Athens, GA. Audiffren is in the Research Institute on Cognition and Learning, UMR CNRS 7295, Sport Sciences Faculty at the University of Poitiers, Poitiers, France. Address author correspondence to Phillip D. Tomporowski at [email protected].
While most studies provide support for a “posture-first” hypothesis, the pattern of dual-task interference has not been consistent and it has been argued that variations in motor movement tasks, cognitive test type, and task instructions contribute to the magnitude of dual-task interference (Bock, 2008). Of particular interest are studies reporting improvements in cognitive performance during dual-task conditions compared with single-task conditions. YogevSeligmann et al. (2010) evaluated the role of instructions given to young and older adults to prioritize walking- and mental-task performance under single- and dual-task conditions. Under the single-task condition, while seated, participants performed a verbal fluency task, which involved recalling as many words as possible from a letter-defined category during 1 min. Participants’ walked at a preferred pace for 1 min under four different conditions: singletask walking alone, dual-task with no prioritization of either task, instruction to give priority to gait, and instructions to give priority to the mental task. Older adults showed greater gait variability during dual-task conditions compared with young adults; however, young and older adults’ verbal fluency performance during walking conditions was superior compared with the seated condition. Schaefer, Lovden, Wieckhorst, and Lindenberger (2010) examined children and young adults’ working memory performance and treadmillwalking gait under single- and dual-task conditions. Under the single-task condition, while seated, participants completed two administrations of an N-back test that consisted of four conditions (1-, 2-, 3-, and 4-back). The N-back test was performed while participants were treadmill walking at a preferred speed (adult mean = 3.73 kph; children mean = 3.09 kph) and again at a fixed lower speed (2.5 kph). Compared with young adults, children’s stride variability increased when working-memory demand was increased. However, the cognitive performance of children and young adults improved while walking at a preferred speed, but not the slower fixed speed, compared with sitting in a chair. Drawing on research that has examined the effects of acute bouts of aerobic exercise on cognitive function (McMorris, Tomporowski & Audiffren, 2009), Yogev-Seligmann et al. (2010) explained the observed dual-task facilitation in terms of physiological arousal. Schaefer et al. (2010) posited two alternative explanations for dual-task improvements in 557
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
558 Tomporowski and Audiffren
cognitive performance: (1) walking-induced physiological arousal leads to increases in performance, or (2) walking at nonpreferred speeds is attentionally demanding and results in degraded cognitive performance when walking at either slower or faster than preferred speeds. We tested these competing explanations by assessing the cognitive performance of young and older adults while they walked at a preferred and at a faster speed. Facilitation of cognitive performance while walking at speeds faster than preferred would support predictions drawn from arousal theory; whereas degraded performance would support resource competition theory. Given the results of many studies that have examined dual-task performance, participants’ mental performance during concurrent walking and cognitive test conditions was predicted to be degraded compared with a single-task condition and older adults’ performance would be compromised more than young adults. Participants’ dual-task performance was also predicted to vary as a function of their level of cardiorespiratory fitness and history of physical activity. Cross-sectional studies and experiments report that the cognitive performance of physically fit individuals is superior to those who are less fit (Hillman, Erickson, & Kramer, 2008). These cognitive benefits have been explained in terms of fitness-related neurological integrity (Churchill et al., 2002) and, as such, are of particular interest to gerontologists interested in conditions that prevent or offset age-related declines in cognitive function. The principal aim of the current study was to examine the cognitive performance in younger and older adults while standing or walking on a treadmill. It was hypothesized that older adults and inactive participants require more resources to control balance while walking than their respective counterparts. We also assumed that walking at a nonpreferred speed would require more executive resources than walking at a preferred speed.
and with the use of flyers and newspaper advertisements. To be included in the study, each participant had to provide a certificate of noncontraindication to physical activity that was authorized by a physician after a medical screening. Another inclusion criterion was to be aged from 18 to 24 for young adults and from 65 to 80 for older adults. Participants were assigned to four experimental groups according to their age and their level of physical activity (PA) assessed through the Historical Leisure Activity Questionnaire (HLAQ; Kriska et al., 1988) and a cardiorespiratory fitness test. Characteristics of these 61 voluntary participants are described in Table 1. Each participant received €30 for their participation. Informed consent was obtained from all participants before data collection. No adverse exercise-related events occurred during the investigation. Three participants did not complete the study; two because of time commitments and the third due to perceived difficulty of testing conditions.
Protocol All the participants carried out three experimental sessions. During the first session, young adults performed the VAMEVAL test (Cazorla & Léger, 1993), while older adults performed the Rockport one-mile test (Kline et al., 1987); these two tests enabled us to estimate the cardiorespiratory fitness (VO2max). The VAMEVAL is a continuous running test with progressive intensity in which the participant exercised to the limit of tolerance. Starting speed is 8.5 kph, and speed is increased 0.5 km every minute. This test is generally used to determine the maximal aerobic speed (MAS) and to estimate the maximal oxygen uptake (VO2max) in athletic clubs. During the second session, participants filled out the HLAQ and the Mini-Mental State Examination (MMSE; Folstein, Folstein, & McHugh, 1975), learned to walk on a treadmill at various speeds, and were then familiarized with the cognitive task. The preferred speed of walking was determined during this second session. During the third and last session, participants performed the cognitive task three times on the treadmill: while standing, while walking 5 min at their preferred speed, or walking at a speed 50% faster than the preferred speed. To isolate the effects of physiological arousal on test performance, the standing condition was always performed first and followed by the two walking conditions, which were counterbalanced.
Method Participants Thirty-one young adults (mean age = 20.81 years, SD = 1.45) and 30 older adults (mean age = 71.53 years, SD = 4.14) took part in this experiment. Young adults were recruited among undergraduate students from the University of Poitiers. Older adults were recruited from senior community centers, civic groups, and fitness centers,
Table 1 Characteristics of the Participants Inactive Young Adults Group A
Active Young Adults Group B
Inactive Older Adults Group C
Active Older Adults Group D
n
16
15
15
15
Gender (M/F)
8/8
8/7
6/9
Variable
Age (years)
Group A vs. Group B
Group C vs. Group D
Group A-B vs. Group C-D
9/6
p = .85
p = .27
p = .90
21.31 (1.25)
20.27 (1.49)
72.27 (3.84)
70.80 (4.43)
p < .05
p = .34
p < .0001
(kg/m2)
22.65 (3.79)
23.97 (2.45)
26.83 (4.31)
23.75 (1.84)
p = .26
p < .05
p < .05
MMSE (max = 30)
29.25 (0.68)
29.27 (0.80)
29.07 (0.88)
29.20 (1.32)
p = .95
p = .75
p = .61
Preferred speed (km/hr)
4.13 (0.76)
4.14 (0.35)
3.67 (0.51)
3.92 (0.57)
p = .97
p = .21
p < .05
HLAQ (MET-hr/week)
14.63 (14.22)
73.99 (39.06)
18.99 (16.64)
31.56 (17.26)
p < .001
p = .05
p < .05
VO2max (mL/min/kg)
39.01 (11.28)
48.63 (5.81)
22.07 (3.80)
31.32 (5.27)
p < .01
p < .001
p < .0001
BMI
Note. BMI = body mass index; MMSE = Mini-Mental State Examination; HLAQ = Historical Leisure Activity Questionnaire.
Dual-task Performance 559
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Determination of Walking Speeds First, participants were familiarized with walking on the treadmill. They were instructed to walk on a treadmill with adjustable speed (JOG S.300; IMBERNON Enterprise, Villeurbanne, France) set with a 3% slope. The speed of the treadmill was progressively increased from 1.5 kph to 7 kph by steps of 0.5 kph every 30 s. This familiarization period lasted 6 min. Participants then performed six trials to determine their preferred speed of walking on the treadmill. Each trial took place in the same way: the speed of the treadmill was set at 2 kph, the participant was then asked to put their feet on the treadmill, and the speed of the treadmill was increased based on the participant’s feedback until a preferred speed was reached. The first trial was considered a familiarization trial and was not taken into account in the computation of preferred speed. Preferred speed was defined as the average of selected speeds of trials 2–6. The faster pace was defined as 50% faster than preferred speed. If the faster pace led the participant to transition from walking to running, the speed was decreased by 10%.
Cognitive Task The cognitive task was an auditory switch test in which 60 computer-generated letters or numbers were presented binaurally to a headphone via a commercial software program. The letters consisted of five vowels (A, E, I, O, and U) and five randomly selected consonants (B, D, L, C, and J). The numbers consisted of four even numbers (2, 4, 6, and 8) and four odd numbers (1, 3, 5, and 7). The participant responded to each stimulus by pressing a key on a serial mouse (even number-left key; odd number-right key; vowel letterleft key; consonant letter-right key). Each key press was followed 100 ms later by the presentation of the next stimulus. Two block types were used: in homogenous blocks, participants responded to letters only or numbers only; in mixed or heterogeneous blocks, participants responded alternatively to letters and numbers. In mixed blocks, letters or numbers were presented in series lengths of one, two, three, or four stimuli. Series lengths were predetermined and counterbalanced. The letter-number category discrimination switched following each series. In each block of trials, the initial 4 trials were considered practice and not evaluated. The remaining trials consisted of nonswitch trials (i.e., repetitive within-category discriminations) and switch trials (i.e., a change in category discrimination). There was an equal number of switches to even-odd and vowel-consonant conditions. Response times (RT) and response accuracy were recorded for each trial. This cognitive task provided a measure of cognitive flexibility, a well-known executive function (Miyake et al., 2000).
Dependent Variables The main dependent variables were mean correct RT, local switch cost (mean RT for switches minus mean RT for repetitions within mixed blocks), mixed switch cost (mean RT for repetitions in homogenous blocks minus mean RT for repetitions in mixed blocks), and decision error rate. Mean RT for switches in mixed blocks was computed as follows: (1) excluding the four first trials of the blocks considered as warm-up trials, (2) excluding any trial corresponding to a decision error, (3) excluding any trial following a decision error, (4) excluding RT < 200 ms and > 3,000 ms, (5) selecting only trials corresponding to a switch from letter to number or the reverse, and (6) averaging selected RTs (there was a maximum of 18 switch trials in each block). Mean RT for repetitions in mixed blocks was computed in the same way except for step 5: the last repetition trial
in a series of letter repetitions or a series of number repetitions was selected. Decision errors were defined as trials during which participants pressed the wrong response key.
Statistical Analysis Univariate analyses of variance (ANOVA) were performed to evaluate the effects of age and level of physical activity on RT time, decision error rate, and switching costs. Repeated-measure multivariate analyses of variance (MANOVA) were conducted instead of ANOVA when testing the interaction of a repeated-measures factor with more than two levels (e.g., test condition) with other factors to avoid any possible problem of sphericity (O’Brien & Kaiser, 1985; Rogan, Keselman, & Mendoza, 1979). When a three-way interaction was significant, planned comparisons were conducted to interpret the higher-order interaction. Post hoc Newman-Keuls tests were also conducted to explain complex interactions. Differences were considered significant at p < .05. When a difference reached significance, effect size was calculated via Cohen’s d or partial eta square (ηp2).
Results First, an ANOVA was conducted on mean RT with age (young adults vs. older adults) and level of PA (active vs. inactive participants) as between-subject factors, and test condition (standing, preferred pace, faster pace), block of trials (1–3), and type of trial (switch vs. repetition) as repeated measure factors. The effect of level of PA did not reach significance, nor did interactions between level of PA and the others factors manipulated in this experiment. By contrast, the three-way interaction among age, test condition, and type of trials reached significance, F(2, 56) = 3.62, p < .05, Wilks’s lambda = 0.89, ηp2 = 11 (see Figure 1). Planned comparisons showed that the interaction of age × type of trials reached significance in older adults, F(2, 56) = 6.57, p < .05, Wilks’s lambda = 0.81, ηp2 = .19, but not in young adults, F(2, 56) = 1.14, p > .32. Post hoc comparisons showed that in young adults the difference in RT between repetitions and switches reached significance but was relatively small and stable (mean delta = 38 ms, mean Cohen’s d = 0.29) in the three test conditions. In contrast, RTs for older adults were significantly slower while standing up than walking at preferred speed and at the higher speed. In addition, the difference between repetitions and switches was larger in the standing condition (M = 196 ms, d = 0.55) than in the two walking conditions (M = 129 ms, d = 0.48). The difference between repetitions and alternations did not differ significantly in the two conditions of walking. A second ANOVA was conducted with the same design on the number of decision errors. The three-way interaction among age, test condition, and type of trial also reached significance, F(2, 56) = 3.34, p < .05, Wilks’s lambda = 0.89, ηp2 = .11 (see Figure 2). Planned comparisons showed that the two-way interaction between test condition and type of trial did not reach significance in young adults, F(2, 56) = 1.03, p > .36, Wilks’s lambda = 0.96, but was significant in older adults, F(2, 56) = 4.02; p < .05, Wilks’s lambda = 0.87, ηp2 = .13. Post hoc tests showed that switches led to more decision errors than repetitions in all test conditions in older adults. However, the effect size of the difference between repetitions and switches increased from standing (d = 0.45), to walking at preferred speed (d = 0.61), and from this last condition to walking at a speed 50% faster than the preferred speed (d = 0.75). Finally, decision error rate for switches was significantly lower while standing (M =
560 Tomporowski and Audiffren
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Figure 1 — Response time (RT) as a function of age of participants (young adults and older adults), test condition (standing, walking at preferred pace, walking at a faster pace 50% higher than the preferred speed), and type of trial (switch and repetition). Error bars represent 95% confidence interval.
Figure 2 — Decision error rate as a function of age of participants (young adults and older adults), test condition (standing, walking at preferred pace, walking at a faster pace 50% higher than the preferred speed), and type of trial (switch and repetition). Error bars represent 95% confidence interval.
6.60%) than walking at preferred speed (M = 8.02%), and significantly higher while walking at a faster pace (M = 10.67%) than at preferred speed. By contrast, for repetitions, there was no significant difference between standing and walking at preferred speed and between the two conditions of walking, but a significant difference between standing (M = 4.30%) and walking at a speed 50% higher than the preferred speed (M = 6.37%). Combining RT and error rate data suggests a speed-accuracy tradeoff for older adults when comparing standing to the two walking conditions. Older adults were slower but more accurate while standing up than while walking.
A third ANOVA was conducted on local switch cost with the same design as the previous analysis. We observed a significant interaction between age and test condition, F(2, 56) = 3.62, p < .05, Wilks’s lambda = 0.89, ηp2 = .11. Post hoc comparisons showed that local switch cost did not vary significantly for young adults as a function of test condition, whereas for older adults switch costs were significantly higher when standing than while walking. A final ANOVA was conducted on mixed switch cost with the same design. The same interaction observed in previous analyses was detected, F(2, 56) = 4.53, p < .05, Wilks’s lambda = 0.86, ηp2 = .14 (see Figure 3).
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Dual-task Performance 561
Figure 3 — Mixed switch cost as a function of age (young adults vs. older adults) and test condition (standing, walking at preferred pace, walking at a faster pace 50% higher than the preferred speed). Mixed switch cost was calculated as the mean RT for repetitions in homogenous blocks minus the mean RT for repetitions in mixed blocks. Error bars represent 95% confidence interval.
Discussion Dual-task experiments that combine walking with mental test conditions typically report degraded cognitive performance, and the effect is particularly evident in studies conducted with older adults. A few studies, however, report the facilitation of participants’ cognitive performance while walking (Decker, Cignetti, Potter, Studenski, & Stergiou, 2012; Schaefer et al., 2010; Yogev-Seligmann et al., 2010). The present study was conducted to address the inconsistencies in these research outcomes by assessing the cognitive performance of young and older adults while they walked on a treadmill at a preferred speed and at a faster than preferred speed. Conforming to predictions from arousal theory (Sanders, 1998), the faster response times and reduced switch costs observed in the current study support the view that cognitive performance may improve under dual-task conditions that involve locomotion and speed of mental processing. When older adults walked at their preferred speed and at an imposed faster pace, we observed significantly faster choice-response times and reductions of both localand mixed-switch costs compared with cognitive test performance measured while standing. Younger adults’ evidenced a similar but nonsignificant reduction in switch-task performance. While choiceresponse speed increased under dual-task conditions, for older adults it was accompanied by a test-specific increase in response errors. Increases in error rates were not observed during test conditions in which participants made repetitive within-category discriminations, but errors did increase on switch trials when participants were required to inhibit a previously used decision-making strategy (e.g., letter-type discrimination) and employ a different decision-making strategy (e.g., a number-type discrimination). The frequency of errors made by older adults increased in a linear fashion across test conditions during which participants stood, walked at a preferred speed, and walked at a faster speed.
Dual-task facilitation in mental performance seen in prior dual-task studies has been previously explained in terms of walkinginduced arousal (Schaefer et al., 2010). Two competing versions of arousal theory have been proposed. The classic 1908 Yerkes-Dodson interpretation predicts an inverted-U relation, with best performance at moderate levels of arousal and poorer performance at low and high levels of arousal. The drive interpretation predicts improved performance with increasing levels of arousal (Landers, 1980). However, in the present experiment, participants’ performance measured in terms of RT or switch costs did not change appreciably when required to walk at a faster than preferred speed. Thus, neither the inverted U-shape nor drive interpretation adequately explains the data obtained. While it is the case that researchers have linked exercise-induced arousal to cognitive function, these changes are typically detected following considerably long (about 20 min) periods of relatively high levels of aerobic exercise demand (McMorris, Sproule, Turner, & Hale, 2011). The facilitation of switch-task performance observed in the current study occurred during the first few minutes of a walking protocol that increased heart rate and required energy utilization levels far less than typical exercise studies. It may be, as suggested by Sanders’ (1983) cognitive-energetic model, that transitory increases in arousal produced by subtle changes in locomotion can alter human information processing and decision making and lead to faster responding (execution of the automatized response) and higher error rate (in some cases, the automatized response is not appropriate). The resource explanation presented by Schaefer et al. (2010) was formulated based on their observation that young adults’ and children’s working memory performance was facilitated when walking at a preferred speed, but not at a fixed slower place. The differences in performance were explained in terms of attentional demands required when walking at a nonpreferred pace. For each individual in the current study, the prescribed walking speed was
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
562 Tomporowski and Audiffren
50% faster than their preferred speed. Given that the increased walking speed was attentionally demanding, the increased pace was predicted to result in less efficient switch-task performance. Examination of error rates revealed that only older adults’ errors increased with walking speed. Thus, while arousal produced by locomotion facilitated older adults’ information processing and lowered response times, it also led to more frequent response errors. The speed-accuracy tradeoff evident in the results of the current study support resource interpretations of attentional allocation. Support for arousal or resource explanations for dual-task interference or facilitation may depend on the cognitive tests selected and the role of speed-accuracy tradeoffs on the quality of performance. The switch task purportedly measures only one aspect of executive functioning–mental flexibility. It remains to be determined whether different dual-task outcomes are observed when using tests that provide measures of updating (working memory) or inhibition. Indeed, the results obtained by Yogev-Seligmann et al. (2010) provide compelling evidence for dual-task facilitation when older adults performed a working memory task. Because the sensitivity of executive tasks to dual-tasking during exercise or physical activity level seems very heterogeneous (Abou-Dest, Albinet, Boucard, & Audiffren, 2012; Boucard et al., 2012) and depends on the choice of the dependent variable (Audiffren, Tomporowski, & Zagrodnik, 2009), it is recommended for further studies to use several indexes of executive performance in the same experiment to explore this multifaceted function. Participants’ level of physical fitness did not influence their dual-task performance. Several studies have found that the processing speeds of older adults who engage in routine exercise are more rapid than less physically active peers (Pesce, Cereatti, Casella, Baldari, & Capranica, 2007; Spirduso, 1975). In the current study, a stringent classification method was employed that took into account both measures of individual participants’ cardiorespiratory fitness and reports of their physical activity history. It was expected that physically active older adults would be particularly at an advantage when performing under challenging dual-task conditions. Additional research will be required to determine the respective role of aerobic physical fitness and physical activity level on dual-task performance. The results of the current study highlight the challenges faced when attempting to identify environmental conditions that may influence older adults’ cognitive function as they move and negotiate complex environments. As discussed by Damos (1991), the characteristics of tasks and their combinations are critical to establishing dual-task interference and decrements in performance. Our observation that older adults’ response accuracy decreased only on switch-task trials reinforces the view that the magnitude of dual-task interference depends on specific processing demands. When walking, older adults may be particularly susceptible to environmental conditions that require rapid changes in strategy selection and utilization. Acknowledgments This research was supported in part by grants from the Regional Council of Poitou-Charentes (France). The authors would like to acknowledge Raphaëlle Château for the recruitment of participants and collecting data.
References Abou-Dest, A., Albinet, C.T., Boucard, G., & Audiffren, M. (2012). Swimming as a positive moderator of cognitive aging: a cross-sectional
study with a multi-task approach. Journal of Aging Research. Article ID, 273185, doi:10.1155/2012/273185. Audiffren, M., Tomporowski, P., & Zagrodnik, J. (2009). Acute aerobic exercise and information processing: Modulation of executive control in a Random Number Generation task. Acta Psychologica, 132, 85–95. PubMed doi:10.1016/j.actpsy.2009.06.008 Bock, O. (2008). Dual-task costs while walking increase in old age for some, but not for other tasks: an experimental study of healthy young and elderly persons. Journal of Neuroengineering and Rehabilitation, 5, 27. doi:10.1186/1743-0003-5-27. PubMed Boucard, G.K., Albinet, C.T., Bugaiska, A., Bouquet, C.A., Clarys, D., & Audiffren, M. (2012). Impact of physical activity on executive functions in aging: A selective effect on inhibition among old adults. Journal of Sport & Exercise Psychology, 34, 808–827. PubMed Cazorla, G., & Léger, L. (1993). Comment évaluer et développer vos capacités aérobies. Epreuves de course navette et épreuve Vam-éval. Bordeaux: Editions AREAPS. Churchill, J.D., Galvez, R., Colcombe, S., Swain, R.A., Kramer, A.F., & Greenough, W.T. (2002). Exercise, experience and the aging brain. Neurobiology of Aging, 23(5), 941–955. PubMed doi:10.1016/S01974580(02)00028-3 Damos, D.L. (1991). Dual-task methodology: Some common problem. In D.L. Damos (Ed.), Multiple-task performance (pp. 101–116). Washington, DC: Taylor & Francis. Decker, L.M., Cignetti, F., Potter, J.F., Studenski, S.A., & Stergiou, N. (2012). Use of motor abundance in young and older adults during dual-task treadmill walking. PLoS ONE, 7(7), e41306. PubMed doi:10.1371/journal.pone.0041306 Folstein, M.F., Folstein, S.E., & McHugh, P.R. (1975). “Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. PubMed doi:10.1016/0022-3956(75)90026-6 Hillman, C.H., Erickson, K.I., & Kramer, A.F. (2008). Be smart, exercise your heart: exercise effects on brain and cognition. Nature Reviews. Neuroscience, 9(1), 58–65. PubMed doi:10.1038/nrn2298 Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: PrenticeHall, Inc. Kline, G.M., Porcari, J.P., Hintermeister, R., Freedson, P.S., Ward, A., McCarron, R.F., . . .Rippe, J.M. (1987). Estimation of VO2max from a one-mile track walk, gender, age, and body weight. Medicine and Science in Sports and Exercise, 19(3), 253–259. PubMed doi:10.1249/00005768-198706000-00012 Kriska, A.M., Sandler, R.B., Cauley, J.A., LaPorte, R.E., Hom, D.L., & Pambianco, G. (1988). The assessment of historical physical activity and its relation to adult bone parameters. American Journal of Epidemiology, 127, 1053–1063. PubMed Landers, D.M. (1980). The arousal-performance relationship revisited. Research Quarterly for Exercise and Sport, 51(1), 77–90. PubMed doi:10.1080/02701367.1980.10609276 Li, K.Z., & Lindenberger, U. (2002). Relations between aging sensory/ sensorimotor and cognitive functions. Neuroscience and Biobehavioral Reviews, 26, 777–783. PubMed doi:10.1016/S0149-7634(02)00073-8 Li, K.Z., Lindenberger, U., Freund, A.M., & Baltes, P.B. (2001). Walking while memorizing: age-related differences in compensatory behavior. Psychological Science, 12(3), 230–237. PubMed doi:10.1111/14679280.00341 Lindenberger, U., Marsiske, M., & Baltes, P.B. (2000). Memorizing while walking: increase in dual-task costs from young adulthood to old age. Psychology and Aging, 15(3), 417–436. PubMed doi:10.1037/08827974.15.3.417 McMorris, T., Sproule, J., Turner, A., & Hale, B. (2011). Acute, intermediate intensity exercise, and speed and accuracy in working memory tasks:
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Dual-task Performance 563 a meta-analytical comparison of effects. Physiology & Behavior, 102, 421–428. PubMed doi:10.1016/j.physbeh.2010.12.007 McMorris, T., Tomporowski, P., & Audiffren, M. (2009). Exercise and cognitive function. Oxford, UK: John Wiley & Sons. Miyake, A., Friedman, N.P., Emerson, M.J., Witzki, A.H., Howerter, A., & Wager, T.D. (2000). The unity and diversity of executive functions and their contributions to complex “Frontal Lobe” tasks: A latent variable analysis. Cognitive Psychology, 41, 49–100. PubMed doi:10.1006/ cogp.1999.0734 Navon, D., & Gopher, D. (1980). Task difficulty, resources, and dual-task performance. In R.S. Nickerson (Ed.), Attention and performance VIII (pp. 297–315). Hillsdale, NJ: Erlbaum. O’Brien, R., & Kaiser, M. (1985). MANOVA method for analyzing repeated measures designs: An extensive primer. Psychological Bulletin, 97, 316–333. PubMed doi:10.1037/0033-2909.97.2.316 Pesce, C., Cereatti, L., Casella, R., Baldari, C., & Capranica, L. (2007). Preservation of visual attention in older expert orienteers at rest and under physical effort. Journal of Sport & Exercise Psychology, 29, 78–99. PubMed Rapp, M.A., Krampe, R.T., & Baltes, P.B. (2006). Adaptive task prioritization in aging: Selective resource allocation to postural control is preserved in Alzheimer Disease. The American Journal of Geriatric Psychiatry, 14(1), 52–61. PubMed doi:10.1097/01. JGP.0000192490.43179.e7 Rogan, J.C., Keselman, H.J., & Mendoza, J.L. (1979). Analysis of repeated measurements. The British Journal of Mathematical and Statistical Psychology, 32, 269–286. doi:10.1111/j.2044-8317.1979.tb00598.x Sanders, A.F. (1983). Towards a model of stress and human performance. Acta Psychologica, 53, 61–97. PubMed doi:10.1016/00016918(83)90016-1
Sanders, A.F. (1998). Elements of human performance. Mahwah, NJ: Lawrence Erlbaum. Schaefer, S., Lovden, M., Wieckhorst, B., & Lindenberger, U. (2010). Cognitive performance is improved while walking: Differences in cognitive-sensorimotor couplings between children and young adults. European Journal of Developmental Psychology, 7(3), 371–389. doi:10.1080/17405620802535666 Shumway-Cook, A., Baldwin, M., Polissar, N.L., & Gruber, W. (1997). Predicting the probability for falls in community-dwelling older adults. Physical Therapy, 77, 812–819. PubMed Spirduso, W.W. (1975). Reaction time and movement time as a function of age and physical activity level. Journal of Gerontology, 30, 435–440. PubMed doi:10.1093/geronj/30.4.435 Wickens, C.D. (2008). Multiple resources and mental workload. Human Factors, 50(3), 449–455. PubMed doi:10.1518/001872008X288394 Woollacott, M., & Shumway-Cook, A.M. (2002). Attention and control of posture and gait: A review of an emerging area of research. Gait & Posture, 16, 1–14. PubMed doi:10.1016/S0966-6362(01)00156-4 Yogev-Seligmann, G., Hausdorff, J.M., & Giladi, N. (2008). The role of executive function and attention in gait. Movement Disorders, 23(3), 329–342. PubMed doi:10.1002/mds.21720 Yogev-Seligmann, G., Rotem-Galili, Y., Mirelman, A., Dickstein, R., Giladi, N., & Hausdorff, J.M. (2010). How does explicit prioritization alter walking during dual-task performance? Effects of age and sex on gait speed and variability. Physical Therapy, 90, 177–186. PubMed doi:10.2522/ptj.20090043
Official Journal of ICAPA www.JAPA-Journal.com ORIGINAL RESEARCH
Dual-task Performance in Young and Older Adults: Speed-Accuracy Tradeoffs in Choice Responding While Treadmill Walking Phillip D. Tomporowski and Michel Audiffren Thirty-one young (mean age = 20.8 years) and 30 older (mean age = 71.5 years) men and women categorized as physically active (n = 30) or inactive (n = 31) performed an executive processing task while standing, treadmill walking at a preferred pace, and treadmill walking at a faster pace. Dual-task interference was predicted to negatively impact older adults’ cognitive flexibility as measured by an auditory switch task more than younger adults; further, participants’ level of physical activity was predicted to mitigate the relation. For older adults, treadmill walking was accompanied by significantly more rapid response times and reductions in local- and mixed-switch costs. A speed-accuracy tradeoff was observed in which response errors increased linearly as walking speed increased, suggesting that locomotion under dual-task conditions degrades the quality of older adults’ cognitive flexibility. Participants’ level of physical activity did not influence cognitive test performance. Keywords: switch task, cognition, dual-task interference, physical activity, information processing Daily activities require the ability to multitask. Researchers in many fields have long been interested in understanding how people perform multiple tasks concurrently. Numerous studies have shown that dual-task demands lead to degraded performance on one or both tasks (see review by Wickens, 2008). These changes in performance are typically explained in terms of competition for limited resources (Kahneman, 1973). The general notion is that the resources required for processing at any point in time are available only in fixed quantities and that performance deteriorates when resources are exceeded by task demands (Navon & Gopher, 1980). Theorists debate whether resources are best characterized as a single, general purpose entity or as multiple pools (Wickens, 2008). Researchers have investigated dual-task conditions that lead to interference between control of balance or locomotion and cognitive processing (Bock, 2008). The dual-task demands of motoric and cognitive processing have been of particular interest to gerontologists, who posit that age-related declines in sensorimotor processes increase the magnitude of task interference (Li & Lindenberger, 2002; Li, Lindenberger, Freund, & Baltes, 2001; Woollacott & Shumway-Cook, 2002). Older adults consistently have been found to prioritize balance (Shumway-Cook, Baldwin, Polissar, & Gruber, 1997) and locomotor control (Lindenberger, Marsiske, & Baltes, 2000; Rapp, Krampe, & Baltes, 2006) over cognitive performance. Data obtained from these studies, and others, have led some to question the traditional view that balance and locomotion control is primarily reflexive and highly automatized and to suggest that higher-level executive functions play a role in these motoric behaviors (Yogev-Seligmann, Hausdorff, & Giladi, 2008). Regardless of the theoretical view, the allocation of resources to control locomotor behaviors is predicted to reduce available resources for processing competing tasks. Tomporowski is in the Department of Kinesiology at the University of Georgia, Athens, GA. Audiffren is in the Research Institute on Cognition and Learning, UMR CNRS 7295, Sport Sciences Faculty at the University of Poitiers, Poitiers, France. Address author correspondence to Phillip D. Tomporowski at [email protected].
While most studies provide support for a “posture-first” hypothesis, the pattern of dual-task interference has not been consistent and it has been argued that variations in motor movement tasks, cognitive test type, and task instructions contribute to the magnitude of dual-task interference (Bock, 2008). Of particular interest are studies reporting improvements in cognitive performance during dual-task conditions compared with single-task conditions. YogevSeligmann et al. (2010) evaluated the role of instructions given to young and older adults to prioritize walking- and mental-task performance under single- and dual-task conditions. Under the single-task condition, while seated, participants performed a verbal fluency task, which involved recalling as many words as possible from a letter-defined category during 1 min. Participants’ walked at a preferred pace for 1 min under four different conditions: singletask walking alone, dual-task with no prioritization of either task, instruction to give priority to gait, and instructions to give priority to the mental task. Older adults showed greater gait variability during dual-task conditions compared with young adults; however, young and older adults’ verbal fluency performance during walking conditions was superior compared with the seated condition. Schaefer, Lovden, Wieckhorst, and Lindenberger (2010) examined children and young adults’ working memory performance and treadmillwalking gait under single- and dual-task conditions. Under the single-task condition, while seated, participants completed two administrations of an N-back test that consisted of four conditions (1-, 2-, 3-, and 4-back). The N-back test was performed while participants were treadmill walking at a preferred speed (adult mean = 3.73 kph; children mean = 3.09 kph) and again at a fixed lower speed (2.5 kph). Compared with young adults, children’s stride variability increased when working-memory demand was increased. However, the cognitive performance of children and young adults improved while walking at a preferred speed, but not the slower fixed speed, compared with sitting in a chair. Drawing on research that has examined the effects of acute bouts of aerobic exercise on cognitive function (McMorris, Tomporowski & Audiffren, 2009), Yogev-Seligmann et al. (2010) explained the observed dual-task facilitation in terms of physiological arousal. Schaefer et al. (2010) posited two alternative explanations for dual-task improvements in 557
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
558 Tomporowski and Audiffren
cognitive performance: (1) walking-induced physiological arousal leads to increases in performance, or (2) walking at nonpreferred speeds is attentionally demanding and results in degraded cognitive performance when walking at either slower or faster than preferred speeds. We tested these competing explanations by assessing the cognitive performance of young and older adults while they walked at a preferred and at a faster speed. Facilitation of cognitive performance while walking at speeds faster than preferred would support predictions drawn from arousal theory; whereas degraded performance would support resource competition theory. Given the results of many studies that have examined dual-task performance, participants’ mental performance during concurrent walking and cognitive test conditions was predicted to be degraded compared with a single-task condition and older adults’ performance would be compromised more than young adults. Participants’ dual-task performance was also predicted to vary as a function of their level of cardiorespiratory fitness and history of physical activity. Cross-sectional studies and experiments report that the cognitive performance of physically fit individuals is superior to those who are less fit (Hillman, Erickson, & Kramer, 2008). These cognitive benefits have been explained in terms of fitness-related neurological integrity (Churchill et al., 2002) and, as such, are of particular interest to gerontologists interested in conditions that prevent or offset age-related declines in cognitive function. The principal aim of the current study was to examine the cognitive performance in younger and older adults while standing or walking on a treadmill. It was hypothesized that older adults and inactive participants require more resources to control balance while walking than their respective counterparts. We also assumed that walking at a nonpreferred speed would require more executive resources than walking at a preferred speed.
and with the use of flyers and newspaper advertisements. To be included in the study, each participant had to provide a certificate of noncontraindication to physical activity that was authorized by a physician after a medical screening. Another inclusion criterion was to be aged from 18 to 24 for young adults and from 65 to 80 for older adults. Participants were assigned to four experimental groups according to their age and their level of physical activity (PA) assessed through the Historical Leisure Activity Questionnaire (HLAQ; Kriska et al., 1988) and a cardiorespiratory fitness test. Characteristics of these 61 voluntary participants are described in Table 1. Each participant received €30 for their participation. Informed consent was obtained from all participants before data collection. No adverse exercise-related events occurred during the investigation. Three participants did not complete the study; two because of time commitments and the third due to perceived difficulty of testing conditions.
Protocol All the participants carried out three experimental sessions. During the first session, young adults performed the VAMEVAL test (Cazorla & Léger, 1993), while older adults performed the Rockport one-mile test (Kline et al., 1987); these two tests enabled us to estimate the cardiorespiratory fitness (VO2max). The VAMEVAL is a continuous running test with progressive intensity in which the participant exercised to the limit of tolerance. Starting speed is 8.5 kph, and speed is increased 0.5 km every minute. This test is generally used to determine the maximal aerobic speed (MAS) and to estimate the maximal oxygen uptake (VO2max) in athletic clubs. During the second session, participants filled out the HLAQ and the Mini-Mental State Examination (MMSE; Folstein, Folstein, & McHugh, 1975), learned to walk on a treadmill at various speeds, and were then familiarized with the cognitive task. The preferred speed of walking was determined during this second session. During the third and last session, participants performed the cognitive task three times on the treadmill: while standing, while walking 5 min at their preferred speed, or walking at a speed 50% faster than the preferred speed. To isolate the effects of physiological arousal on test performance, the standing condition was always performed first and followed by the two walking conditions, which were counterbalanced.
Method Participants Thirty-one young adults (mean age = 20.81 years, SD = 1.45) and 30 older adults (mean age = 71.53 years, SD = 4.14) took part in this experiment. Young adults were recruited among undergraduate students from the University of Poitiers. Older adults were recruited from senior community centers, civic groups, and fitness centers,
Table 1 Characteristics of the Participants Inactive Young Adults Group A
Active Young Adults Group B
Inactive Older Adults Group C
Active Older Adults Group D
n
16
15
15
15
Gender (M/F)
8/8
8/7
6/9
Variable
Age (years)
Group A vs. Group B
Group C vs. Group D
Group A-B vs. Group C-D
9/6
p = .85
p = .27
p = .90
21.31 (1.25)
20.27 (1.49)
72.27 (3.84)
70.80 (4.43)
p < .05
p = .34
p < .0001
(kg/m2)
22.65 (3.79)
23.97 (2.45)
26.83 (4.31)
23.75 (1.84)
p = .26
p < .05
p < .05
MMSE (max = 30)
29.25 (0.68)
29.27 (0.80)
29.07 (0.88)
29.20 (1.32)
p = .95
p = .75
p = .61
Preferred speed (km/hr)
4.13 (0.76)
4.14 (0.35)
3.67 (0.51)
3.92 (0.57)
p = .97
p = .21
p < .05
HLAQ (MET-hr/week)
14.63 (14.22)
73.99 (39.06)
18.99 (16.64)
31.56 (17.26)
p < .001
p = .05
p < .05
VO2max (mL/min/kg)
39.01 (11.28)
48.63 (5.81)
22.07 (3.80)
31.32 (5.27)
p < .01
p < .001
p < .0001
BMI
Note. BMI = body mass index; MMSE = Mini-Mental State Examination; HLAQ = Historical Leisure Activity Questionnaire.
Dual-task Performance 559
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Determination of Walking Speeds First, participants were familiarized with walking on the treadmill. They were instructed to walk on a treadmill with adjustable speed (JOG S.300; IMBERNON Enterprise, Villeurbanne, France) set with a 3% slope. The speed of the treadmill was progressively increased from 1.5 kph to 7 kph by steps of 0.5 kph every 30 s. This familiarization period lasted 6 min. Participants then performed six trials to determine their preferred speed of walking on the treadmill. Each trial took place in the same way: the speed of the treadmill was set at 2 kph, the participant was then asked to put their feet on the treadmill, and the speed of the treadmill was increased based on the participant’s feedback until a preferred speed was reached. The first trial was considered a familiarization trial and was not taken into account in the computation of preferred speed. Preferred speed was defined as the average of selected speeds of trials 2–6. The faster pace was defined as 50% faster than preferred speed. If the faster pace led the participant to transition from walking to running, the speed was decreased by 10%.
Cognitive Task The cognitive task was an auditory switch test in which 60 computer-generated letters or numbers were presented binaurally to a headphone via a commercial software program. The letters consisted of five vowels (A, E, I, O, and U) and five randomly selected consonants (B, D, L, C, and J). The numbers consisted of four even numbers (2, 4, 6, and 8) and four odd numbers (1, 3, 5, and 7). The participant responded to each stimulus by pressing a key on a serial mouse (even number-left key; odd number-right key; vowel letterleft key; consonant letter-right key). Each key press was followed 100 ms later by the presentation of the next stimulus. Two block types were used: in homogenous blocks, participants responded to letters only or numbers only; in mixed or heterogeneous blocks, participants responded alternatively to letters and numbers. In mixed blocks, letters or numbers were presented in series lengths of one, two, three, or four stimuli. Series lengths were predetermined and counterbalanced. The letter-number category discrimination switched following each series. In each block of trials, the initial 4 trials were considered practice and not evaluated. The remaining trials consisted of nonswitch trials (i.e., repetitive within-category discriminations) and switch trials (i.e., a change in category discrimination). There was an equal number of switches to even-odd and vowel-consonant conditions. Response times (RT) and response accuracy were recorded for each trial. This cognitive task provided a measure of cognitive flexibility, a well-known executive function (Miyake et al., 2000).
Dependent Variables The main dependent variables were mean correct RT, local switch cost (mean RT for switches minus mean RT for repetitions within mixed blocks), mixed switch cost (mean RT for repetitions in homogenous blocks minus mean RT for repetitions in mixed blocks), and decision error rate. Mean RT for switches in mixed blocks was computed as follows: (1) excluding the four first trials of the blocks considered as warm-up trials, (2) excluding any trial corresponding to a decision error, (3) excluding any trial following a decision error, (4) excluding RT < 200 ms and > 3,000 ms, (5) selecting only trials corresponding to a switch from letter to number or the reverse, and (6) averaging selected RTs (there was a maximum of 18 switch trials in each block). Mean RT for repetitions in mixed blocks was computed in the same way except for step 5: the last repetition trial
in a series of letter repetitions or a series of number repetitions was selected. Decision errors were defined as trials during which participants pressed the wrong response key.
Statistical Analysis Univariate analyses of variance (ANOVA) were performed to evaluate the effects of age and level of physical activity on RT time, decision error rate, and switching costs. Repeated-measure multivariate analyses of variance (MANOVA) were conducted instead of ANOVA when testing the interaction of a repeated-measures factor with more than two levels (e.g., test condition) with other factors to avoid any possible problem of sphericity (O’Brien & Kaiser, 1985; Rogan, Keselman, & Mendoza, 1979). When a three-way interaction was significant, planned comparisons were conducted to interpret the higher-order interaction. Post hoc Newman-Keuls tests were also conducted to explain complex interactions. Differences were considered significant at p < .05. When a difference reached significance, effect size was calculated via Cohen’s d or partial eta square (ηp2).
Results First, an ANOVA was conducted on mean RT with age (young adults vs. older adults) and level of PA (active vs. inactive participants) as between-subject factors, and test condition (standing, preferred pace, faster pace), block of trials (1–3), and type of trial (switch vs. repetition) as repeated measure factors. The effect of level of PA did not reach significance, nor did interactions between level of PA and the others factors manipulated in this experiment. By contrast, the three-way interaction among age, test condition, and type of trials reached significance, F(2, 56) = 3.62, p < .05, Wilks’s lambda = 0.89, ηp2 = 11 (see Figure 1). Planned comparisons showed that the interaction of age × type of trials reached significance in older adults, F(2, 56) = 6.57, p < .05, Wilks’s lambda = 0.81, ηp2 = .19, but not in young adults, F(2, 56) = 1.14, p > .32. Post hoc comparisons showed that in young adults the difference in RT between repetitions and switches reached significance but was relatively small and stable (mean delta = 38 ms, mean Cohen’s d = 0.29) in the three test conditions. In contrast, RTs for older adults were significantly slower while standing up than walking at preferred speed and at the higher speed. In addition, the difference between repetitions and switches was larger in the standing condition (M = 196 ms, d = 0.55) than in the two walking conditions (M = 129 ms, d = 0.48). The difference between repetitions and alternations did not differ significantly in the two conditions of walking. A second ANOVA was conducted with the same design on the number of decision errors. The three-way interaction among age, test condition, and type of trial also reached significance, F(2, 56) = 3.34, p < .05, Wilks’s lambda = 0.89, ηp2 = .11 (see Figure 2). Planned comparisons showed that the two-way interaction between test condition and type of trial did not reach significance in young adults, F(2, 56) = 1.03, p > .36, Wilks’s lambda = 0.96, but was significant in older adults, F(2, 56) = 4.02; p < .05, Wilks’s lambda = 0.87, ηp2 = .13. Post hoc tests showed that switches led to more decision errors than repetitions in all test conditions in older adults. However, the effect size of the difference between repetitions and switches increased from standing (d = 0.45), to walking at preferred speed (d = 0.61), and from this last condition to walking at a speed 50% faster than the preferred speed (d = 0.75). Finally, decision error rate for switches was significantly lower while standing (M =
560 Tomporowski and Audiffren
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Figure 1 — Response time (RT) as a function of age of participants (young adults and older adults), test condition (standing, walking at preferred pace, walking at a faster pace 50% higher than the preferred speed), and type of trial (switch and repetition). Error bars represent 95% confidence interval.
Figure 2 — Decision error rate as a function of age of participants (young adults and older adults), test condition (standing, walking at preferred pace, walking at a faster pace 50% higher than the preferred speed), and type of trial (switch and repetition). Error bars represent 95% confidence interval.
6.60%) than walking at preferred speed (M = 8.02%), and significantly higher while walking at a faster pace (M = 10.67%) than at preferred speed. By contrast, for repetitions, there was no significant difference between standing and walking at preferred speed and between the two conditions of walking, but a significant difference between standing (M = 4.30%) and walking at a speed 50% higher than the preferred speed (M = 6.37%). Combining RT and error rate data suggests a speed-accuracy tradeoff for older adults when comparing standing to the two walking conditions. Older adults were slower but more accurate while standing up than while walking.
A third ANOVA was conducted on local switch cost with the same design as the previous analysis. We observed a significant interaction between age and test condition, F(2, 56) = 3.62, p < .05, Wilks’s lambda = 0.89, ηp2 = .11. Post hoc comparisons showed that local switch cost did not vary significantly for young adults as a function of test condition, whereas for older adults switch costs were significantly higher when standing than while walking. A final ANOVA was conducted on mixed switch cost with the same design. The same interaction observed in previous analyses was detected, F(2, 56) = 4.53, p < .05, Wilks’s lambda = 0.86, ηp2 = .14 (see Figure 3).
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Dual-task Performance 561
Figure 3 — Mixed switch cost as a function of age (young adults vs. older adults) and test condition (standing, walking at preferred pace, walking at a faster pace 50% higher than the preferred speed). Mixed switch cost was calculated as the mean RT for repetitions in homogenous blocks minus the mean RT for repetitions in mixed blocks. Error bars represent 95% confidence interval.
Discussion Dual-task experiments that combine walking with mental test conditions typically report degraded cognitive performance, and the effect is particularly evident in studies conducted with older adults. A few studies, however, report the facilitation of participants’ cognitive performance while walking (Decker, Cignetti, Potter, Studenski, & Stergiou, 2012; Schaefer et al., 2010; Yogev-Seligmann et al., 2010). The present study was conducted to address the inconsistencies in these research outcomes by assessing the cognitive performance of young and older adults while they walked on a treadmill at a preferred speed and at a faster than preferred speed. Conforming to predictions from arousal theory (Sanders, 1998), the faster response times and reduced switch costs observed in the current study support the view that cognitive performance may improve under dual-task conditions that involve locomotion and speed of mental processing. When older adults walked at their preferred speed and at an imposed faster pace, we observed significantly faster choice-response times and reductions of both localand mixed-switch costs compared with cognitive test performance measured while standing. Younger adults’ evidenced a similar but nonsignificant reduction in switch-task performance. While choiceresponse speed increased under dual-task conditions, for older adults it was accompanied by a test-specific increase in response errors. Increases in error rates were not observed during test conditions in which participants made repetitive within-category discriminations, but errors did increase on switch trials when participants were required to inhibit a previously used decision-making strategy (e.g., letter-type discrimination) and employ a different decision-making strategy (e.g., a number-type discrimination). The frequency of errors made by older adults increased in a linear fashion across test conditions during which participants stood, walked at a preferred speed, and walked at a faster speed.
Dual-task facilitation in mental performance seen in prior dual-task studies has been previously explained in terms of walkinginduced arousal (Schaefer et al., 2010). Two competing versions of arousal theory have been proposed. The classic 1908 Yerkes-Dodson interpretation predicts an inverted-U relation, with best performance at moderate levels of arousal and poorer performance at low and high levels of arousal. The drive interpretation predicts improved performance with increasing levels of arousal (Landers, 1980). However, in the present experiment, participants’ performance measured in terms of RT or switch costs did not change appreciably when required to walk at a faster than preferred speed. Thus, neither the inverted U-shape nor drive interpretation adequately explains the data obtained. While it is the case that researchers have linked exercise-induced arousal to cognitive function, these changes are typically detected following considerably long (about 20 min) periods of relatively high levels of aerobic exercise demand (McMorris, Sproule, Turner, & Hale, 2011). The facilitation of switch-task performance observed in the current study occurred during the first few minutes of a walking protocol that increased heart rate and required energy utilization levels far less than typical exercise studies. It may be, as suggested by Sanders’ (1983) cognitive-energetic model, that transitory increases in arousal produced by subtle changes in locomotion can alter human information processing and decision making and lead to faster responding (execution of the automatized response) and higher error rate (in some cases, the automatized response is not appropriate). The resource explanation presented by Schaefer et al. (2010) was formulated based on their observation that young adults’ and children’s working memory performance was facilitated when walking at a preferred speed, but not at a fixed slower place. The differences in performance were explained in terms of attentional demands required when walking at a nonpreferred pace. For each individual in the current study, the prescribed walking speed was
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
562 Tomporowski and Audiffren
50% faster than their preferred speed. Given that the increased walking speed was attentionally demanding, the increased pace was predicted to result in less efficient switch-task performance. Examination of error rates revealed that only older adults’ errors increased with walking speed. Thus, while arousal produced by locomotion facilitated older adults’ information processing and lowered response times, it also led to more frequent response errors. The speed-accuracy tradeoff evident in the results of the current study support resource interpretations of attentional allocation. Support for arousal or resource explanations for dual-task interference or facilitation may depend on the cognitive tests selected and the role of speed-accuracy tradeoffs on the quality of performance. The switch task purportedly measures only one aspect of executive functioning–mental flexibility. It remains to be determined whether different dual-task outcomes are observed when using tests that provide measures of updating (working memory) or inhibition. Indeed, the results obtained by Yogev-Seligmann et al. (2010) provide compelling evidence for dual-task facilitation when older adults performed a working memory task. Because the sensitivity of executive tasks to dual-tasking during exercise or physical activity level seems very heterogeneous (Abou-Dest, Albinet, Boucard, & Audiffren, 2012; Boucard et al., 2012) and depends on the choice of the dependent variable (Audiffren, Tomporowski, & Zagrodnik, 2009), it is recommended for further studies to use several indexes of executive performance in the same experiment to explore this multifaceted function. Participants’ level of physical fitness did not influence their dual-task performance. Several studies have found that the processing speeds of older adults who engage in routine exercise are more rapid than less physically active peers (Pesce, Cereatti, Casella, Baldari, & Capranica, 2007; Spirduso, 1975). In the current study, a stringent classification method was employed that took into account both measures of individual participants’ cardiorespiratory fitness and reports of their physical activity history. It was expected that physically active older adults would be particularly at an advantage when performing under challenging dual-task conditions. Additional research will be required to determine the respective role of aerobic physical fitness and physical activity level on dual-task performance. The results of the current study highlight the challenges faced when attempting to identify environmental conditions that may influence older adults’ cognitive function as they move and negotiate complex environments. As discussed by Damos (1991), the characteristics of tasks and their combinations are critical to establishing dual-task interference and decrements in performance. Our observation that older adults’ response accuracy decreased only on switch-task trials reinforces the view that the magnitude of dual-task interference depends on specific processing demands. When walking, older adults may be particularly susceptible to environmental conditions that require rapid changes in strategy selection and utilization. Acknowledgments This research was supported in part by grants from the Regional Council of Poitou-Charentes (France). The authors would like to acknowledge Raphaëlle Château for the recruitment of participants and collecting data.
References Abou-Dest, A., Albinet, C.T., Boucard, G., & Audiffren, M. (2012). Swimming as a positive moderator of cognitive aging: a cross-sectional
study with a multi-task approach. Journal of Aging Research. Article ID, 273185, doi:10.1155/2012/273185. Audiffren, M., Tomporowski, P., & Zagrodnik, J. (2009). Acute aerobic exercise and information processing: Modulation of executive control in a Random Number Generation task. Acta Psychologica, 132, 85–95. PubMed doi:10.1016/j.actpsy.2009.06.008 Bock, O. (2008). Dual-task costs while walking increase in old age for some, but not for other tasks: an experimental study of healthy young and elderly persons. Journal of Neuroengineering and Rehabilitation, 5, 27. doi:10.1186/1743-0003-5-27. PubMed Boucard, G.K., Albinet, C.T., Bugaiska, A., Bouquet, C.A., Clarys, D., & Audiffren, M. (2012). Impact of physical activity on executive functions in aging: A selective effect on inhibition among old adults. Journal of Sport & Exercise Psychology, 34, 808–827. PubMed Cazorla, G., & Léger, L. (1993). Comment évaluer et développer vos capacités aérobies. Epreuves de course navette et épreuve Vam-éval. Bordeaux: Editions AREAPS. Churchill, J.D., Galvez, R., Colcombe, S., Swain, R.A., Kramer, A.F., & Greenough, W.T. (2002). Exercise, experience and the aging brain. Neurobiology of Aging, 23(5), 941–955. PubMed doi:10.1016/S01974580(02)00028-3 Damos, D.L. (1991). Dual-task methodology: Some common problem. In D.L. Damos (Ed.), Multiple-task performance (pp. 101–116). Washington, DC: Taylor & Francis. Decker, L.M., Cignetti, F., Potter, J.F., Studenski, S.A., & Stergiou, N. (2012). Use of motor abundance in young and older adults during dual-task treadmill walking. PLoS ONE, 7(7), e41306. PubMed doi:10.1371/journal.pone.0041306 Folstein, M.F., Folstein, S.E., & McHugh, P.R. (1975). “Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. PubMed doi:10.1016/0022-3956(75)90026-6 Hillman, C.H., Erickson, K.I., & Kramer, A.F. (2008). Be smart, exercise your heart: exercise effects on brain and cognition. Nature Reviews. Neuroscience, 9(1), 58–65. PubMed doi:10.1038/nrn2298 Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: PrenticeHall, Inc. Kline, G.M., Porcari, J.P., Hintermeister, R., Freedson, P.S., Ward, A., McCarron, R.F., . . .Rippe, J.M. (1987). Estimation of VO2max from a one-mile track walk, gender, age, and body weight. Medicine and Science in Sports and Exercise, 19(3), 253–259. PubMed doi:10.1249/00005768-198706000-00012 Kriska, A.M., Sandler, R.B., Cauley, J.A., LaPorte, R.E., Hom, D.L., & Pambianco, G. (1988). The assessment of historical physical activity and its relation to adult bone parameters. American Journal of Epidemiology, 127, 1053–1063. PubMed Landers, D.M. (1980). The arousal-performance relationship revisited. Research Quarterly for Exercise and Sport, 51(1), 77–90. PubMed doi:10.1080/02701367.1980.10609276 Li, K.Z., & Lindenberger, U. (2002). Relations between aging sensory/ sensorimotor and cognitive functions. Neuroscience and Biobehavioral Reviews, 26, 777–783. PubMed doi:10.1016/S0149-7634(02)00073-8 Li, K.Z., Lindenberger, U., Freund, A.M., & Baltes, P.B. (2001). Walking while memorizing: age-related differences in compensatory behavior. Psychological Science, 12(3), 230–237. PubMed doi:10.1111/14679280.00341 Lindenberger, U., Marsiske, M., & Baltes, P.B. (2000). Memorizing while walking: increase in dual-task costs from young adulthood to old age. Psychology and Aging, 15(3), 417–436. PubMed doi:10.1037/08827974.15.3.417 McMorris, T., Sproule, J., Turner, A., & Hale, B. (2011). Acute, intermediate intensity exercise, and speed and accuracy in working memory tasks:
Downloaded by Northern Illinois University on 09/19/16, Volume 22, Article Number 4
Dual-task Performance 563 a meta-analytical comparison of effects. Physiology & Behavior, 102, 421–428. PubMed doi:10.1016/j.physbeh.2010.12.007 McMorris, T., Tomporowski, P., & Audiffren, M. (2009). Exercise and cognitive function. Oxford, UK: John Wiley & Sons. Miyake, A., Friedman, N.P., Emerson, M.J., Witzki, A.H., Howerter, A., & Wager, T.D. (2000). The unity and diversity of executive functions and their contributions to complex “Frontal Lobe” tasks: A latent variable analysis. Cognitive Psychology, 41, 49–100. PubMed doi:10.1006/ cogp.1999.0734 Navon, D., & Gopher, D. (1980). Task difficulty, resources, and dual-task performance. In R.S. Nickerson (Ed.), Attention and performance VIII (pp. 297–315). Hillsdale, NJ: Erlbaum. O’Brien, R., & Kaiser, M. (1985). MANOVA method for analyzing repeated measures designs: An extensive primer. Psychological Bulletin, 97, 316–333. PubMed doi:10.1037/0033-2909.97.2.316 Pesce, C., Cereatti, L., Casella, R., Baldari, C., & Capranica, L. (2007). Preservation of visual attention in older expert orienteers at rest and under physical effort. Journal of Sport & Exercise Psychology, 29, 78–99. PubMed Rapp, M.A., Krampe, R.T., & Baltes, P.B. (2006). Adaptive task prioritization in aging: Selective resource allocation to postural control is preserved in Alzheimer Disease. The American Journal of Geriatric Psychiatry, 14(1), 52–61. PubMed doi:10.1097/01. JGP.0000192490.43179.e7 Rogan, J.C., Keselman, H.J., & Mendoza, J.L. (1979). Analysis of repeated measurements. The British Journal of Mathematical and Statistical Psychology, 32, 269–286. doi:10.1111/j.2044-8317.1979.tb00598.x Sanders, A.F. (1983). Towards a model of stress and human performance. Acta Psychologica, 53, 61–97. PubMed doi:10.1016/00016918(83)90016-1
Sanders, A.F. (1998). Elements of human performance. Mahwah, NJ: Lawrence Erlbaum. Schaefer, S., Lovden, M., Wieckhorst, B., & Lindenberger, U. (2010). Cognitive performance is improved while walking: Differences in cognitive-sensorimotor couplings between children and young adults. European Journal of Developmental Psychology, 7(3), 371–389. doi:10.1080/17405620802535666 Shumway-Cook, A., Baldwin, M., Polissar, N.L., & Gruber, W. (1997). Predicting the probability for falls in community-dwelling older adults. Physical Therapy, 77, 812–819. PubMed Spirduso, W.W. (1975). Reaction time and movement time as a function of age and physical activity level. Journal of Gerontology, 30, 435–440. PubMed doi:10.1093/geronj/30.4.435 Wickens, C.D. (2008). Multiple resources and mental workload. Human Factors, 50(3), 449–455. PubMed doi:10.1518/001872008X288394 Woollacott, M., & Shumway-Cook, A.M. (2002). Attention and control of posture and gait: A review of an emerging area of research. Gait & Posture, 16, 1–14. PubMed doi:10.1016/S0966-6362(01)00156-4 Yogev-Seligmann, G., Hausdorff, J.M., & Giladi, N. (2008). The role of executive function and attention in gait. Movement Disorders, 23(3), 329–342. PubMed doi:10.1002/mds.21720 Yogev-Seligmann, G., Rotem-Galili, Y., Mirelman, A., Dickstein, R., Giladi, N., & Hausdorff, J.M. (2010). How does explicit prioritization alter walking during dual-task performance? Effects of age and sex on gait speed and variability. Physical Therapy, 90, 177–186. PubMed doi:10.2522/ptj.20090043
